Establishing cut-points for physical activity classification using triaxial accelerometer in middle-aged recreational...

Area of Science:

• Exercise Physiology
• Biomedical Engineering
• Sports Science

Background:

• Accurate physical activity intensity classification is crucial for understanding energy expenditure in athletes.
• Recreational marathoners represent a key demographic for studying exercise intensity and its effects.
• Existing methods for activity intensity monitoring require further validation with wearable technology.

Purpose of the Study:

• To establish Gravity Estimator of Normal Everyday Activity (GENEA) cut-points for classifying six relative-intensity physical activity levels.
• To validate the accuracy of the GENEA accelerometer in recreational marathoners aged 30-45 years.
• To assess the GENEA device's ability to predict energy expenditure during running activities.

Main Methods:

• Ninety-eight recreational marathoners (83 males, 15 females) underwent treadmill testing with GENEA accelerometers and V̇O2 measurement.
• Receiver Operating Characteristic (ROC) analysis was used to determine GENEA cut-points for activity classification.
• Spearman's correlation assessed the relationship between estimated and measured intensity classifications, analyzed for all participants and by sex.

Main Results:

• GENEA cut-points demonstrated excellent classification accuracy (AUC 0.886–0.973) across all six intensity levels for all participants.
• High classification accuracy was maintained when data was analyzed separately for males (AUC 0.881–0.973) and females (AUC 0.924–0.968).
• GENEA accelerometer data explained a significant variance in energy expenditure (78.50% overall, 78.14% males, 83.17% females).

Conclusions:

• The wrist-worn GENEA accelerometer effectively classifies physical activity intensity in middle-aged recreational marathoners.
• The device shows high accuracy regardless of whether data is analyzed collectively or separated by sex.
• Triaxial GENEA accelerometers show promise for predicting energy expenditure in running activities.